Looking Back and Looking Forward — On Thesis Statements, Prompts, and AI in Education

(Looking Back and Looking Forward takes a look at the articles and posts I found interesting from the previous week, along with reflections about how the trends they point to might shape my thinking about education, technology, and culture.)

At the outset of my academic career, I spent a good bit of time teaching composition courses (in English and Spanish). A cornerstone of those courses was the famed “thesis statement.”

You probably remember this one. A good thesis statement contains both the “what” and the “how/why” related to an essay.

Incomplete Thesis Statement: Internet use has a positive impact on learning.

This statement only addresses the “what” and omits the “why or “how.”

Improved Thesis Statement: The use of the Internet improves the learning experience by connecting students to additional resources and communities.

The statement now contains both a “what” — Internet usage improves the learning experience — AND the “why”/”how.”

While generative AI may be putting the survival of essay writing as a core component of the curriculum in doubt, I think the thesis statement has an important analog in the future of education: the prompt. That’s because we are moving rapidly into a world where a fundamental skill (if not THE fundamental skill) will be the ability to instruct AI tools in the creation of narrative text, images, data, and code. And that skill requires us to teach students, from an early age, how to construct effective prompt statements.

Not surprisingly, thesis statements and generative AI prompts share a number of common elements. The most useful versions of each contain multiple parts or components. In addition, they are both designed to produce illicit responses from an audience (be it in a reader or an LLM) and the quality of those responses correlates in large part fo the context and specificity of the statement or prompt.

The similarities are such, in fact, that I think we should consider abandoning the thesis statement altogether in favor of teaching effective prompt statements. Being language/meaning based, the same people that teach the former can also learn to teach the latter. It’s simply a matter of swapping out the instructions related to what makes a “good” statement and how to identify the different components and their purpose. In the case of prompts, possible elements include:

Instruction: A specific task to be performed by the model.
Context: Additional information so that the model can respond better.
Input: A question that we ask the model.
Indicator/Response Format: Specifies the type of output.
Constraints or Limitations: Set any particular boundaries or limitations the system should consider while generating an output (this could be word count, number of examples, etc.)

And, unlike the case with thesis statements, where feedback generally comes in the form of correction from an instructor, prompts provide immediate results that invite exploration and iterative attempts to get the desired output. That’s right — real-time, helpful feedback and learning.

I see this evolution as both useful and positive in the learning experience. Like Maha Bali, I see generative as “technology the learner can choose to do whatever they want with – rather than technology that teachers/institutions do TO and ABOUT the learner.”

How fast will educators and our education systems adapt to generative AI? It’s hard to say. Donald Clark’s recent experience at the Learning Technologies Conference paints a good picture of the gap between educational technology’s past and future, and George Siemens provides what is likely an accurate assessment of higher education attitudes around AI (as well as what the attitudes should be).

One of the early impacts we’re already seeing is the declining value of information without application or alignment to skills and skill pathways. This past week we witnessed admissions by homework helper Chegg that ChatGPT (free) was responsible for that company declining revenue and subscriber base. That news led to an immediate drop in investor confidence related to the educational publishing giant (read information purveyor) Pearson.

Finally, with all the big tech hype around AI, it’s easy to forget that this particular “arms race” also includes plenty of open-source competitors. As one Google insider puts it, “The uncomfortable truth is, we aren’t positioned to win this arms race and neither is OpenAI. While we’ve been squabbling, a third faction has been quietly eating our lunch. I’m talking, of course, about open source.”

Looking Back and Looking Forward — On Thesis Statements, Prompts, and AI in Education

Further Reading

Higher Education

K-!2 Education

Workforce and Organizational Trends

Online Learning, Learning Design, and Education Technology

Technology, Science, and Culture